Csmd3 for evaluating prognosis of glioma patients + Microglia cell detection method

Abstract

The application relates to the field of biological medicine, and discloses a Csmd3 for evaluating the prognosis of a glioma patient + A microglia cell detection method comprises obtaining clinical transcriptome data and a glioma tissue sample; obtaining a Csmd3 + A target subpopulation characteristic gene set and a total microglia cell characteristic gene set are constructed based on microglia cell characteristic genes; a single-sample gene set enrichment analysis algorithm is configured to calculate subpopulation enrichment scores and total microglia cell enrichment scores of each patient sample, and a ratio of the two is taken as a clinical abundance parameter of the target cell subpopulation; the median value of the ratio parameter is used to divide the patient samples into a high-abundance sample group and a low-abundance sample group; clinical survival period data is input into a survival analysis algorithm for calculation; and a verification result that the clinical abundance parameter is positively correlated with the survival prognosis of the glioma patient is output. By obtaining accurate tumor microenvironment cytology data, the defects of conventional methods disturbed by sequencing noise are eliminated, and a substantial intervention target for accurate prognosis evaluation and immunotherapy of glioma is established.

Description

Technical Field

[0001] This invention relates to the field of biomedical technology, specifically to a Csmd3 for assessing the prognosis of glioma patients. + Microglial cell detection method. Background Technology

[0002] Gliomas are the most common primary malignant tumors of the central nervous system, exhibiting high heterogeneity and invasiveness. Currently, clinical prognostic assessment of glioma patients mainly relies on histopathological grading and a few routine molecular markers. However, the tumor microenvironment of gliomas is complex, and routine detection methods are insufficient to deeply reflect the specific impact of immune microenvironment characteristics on long-term patient survival. Existing prognostic assessment methods based on macroscopic tissues or single markers lack precise microenvironmental cytology evidence, leading to significant individual differences in prognostic prediction results and making it difficult to accurately assess patient survival and provide reliable clinical data support.

[0003] In exploring potential target cells in the tumor immune microenvironment, current technologies largely rely on static clinical samples or conventional primary tumor-bearing animal models. These conventional models can only reflect the microenvironmental state at the time of tumor development or initial intervention, and cannot effectively simulate and reproduce the establishment and evolution of long-term anti-GBM immune memory. Because the industry lacks animal models that can effectively simulate the evolution of "long-term survival" in glioma patients or reproduce the state of "immunocure" after experiencing a "second tumor challenge," current research struggles to reliably identify specific immune cell subsets and their characteristic genes that truly drive and maintain long-term survival.

[0004] Furthermore, existing research on preliminarily screened tumor microenvironment cell populations is often limited to single-dimensional phenotypic correlation analysis or simple basic expression level verification (such as histological fluorescence staining or Western blot detection), lacking a systematic functional intervention and mechanism confirmation paradigm. Although modern cell modification technologies such as gene editing are relatively mature, when applying them to confirm tumor microenvironment targets, existing research generally lacks a comprehensive closed-loop evaluation system covering "in vitro cell activity targeting regulation - in vivo tumor immune microenvironment remodeling assessment - combined immune checkpoint therapy response". This lack of a systematic confirmation mechanism means that the screened cell markers remain at the theoretical correlation stage and are difficult to transform into clear clinical intervention targets, limiting their practical application in the development of novel glioma immunotherapies. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a Csmd3 method for assessing the prognosis of glioma patients. +The microglia detection method addresses the problems of low prediction accuracy due to the lack of precise microenvironmental cytology evidence in existing glioma prognostic assessments, the inability of conventional screening models to stably identify anti-GBM immune memory cell populations due to data noise interference, and the difficulty in substantially transforming preliminary screened cell markers into clinically effective intervention targets due to the lack of systematic in vivo and in vitro functional and combined drug response confirmation.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a Csmd3 for assessing the prognosis of glioma patients. + The method for detecting microglia is characterized by comprising: Acquire clinical transcriptomic data and glioma tissue samples, wherein the clinical transcriptomic data includes transcriptional sequencing data from multiple glioma patient samples; Obtain the characteristic gene set of the target subpopulation and the characteristic gene set of total microglia. The characteristic gene set of the target subpopulation is constructed based on the characteristic genes of the target cell subpopulation, which is Csmd3. + Microglia, wherein the total microglia characteristic gene set is constructed based on the characteristic genes of total microglia; Configure a single-sample gene set enrichment analysis algorithm, and use the single-sample gene set enrichment analysis algorithm to calculate the subgroup enrichment score of each glioma patient sample in the multiple glioma patient samples in the clinical transcriptome data relative to the target subgroup characteristic gene set, and the total microglia enrichment score relative to the total microglia characteristic gene set. The relative proportion value is calculated based on the target subpopulation enrichment score and the total microglia enrichment score, and the relative proportion value is determined as the clinical abundance parameter of the target cell subpopulation in each glioma patient sample. Configure a survival analysis algorithm to calculate the median value of the clinical abundance parameter, and use this as a dividing threshold to divide the multiple glioma patient samples corresponding to the clinical transcriptome data into a high abundance sample group and a low abundance sample group. Acquire clinical survival data, input the high-abundance sample group, the low-abundance sample group, and the clinical survival data into the survival analysis algorithm for calculation, and obtain survival prognosis assessment data; Using the survival prognostic assessment data, the overall survival of the high-abundance sample group and the low-abundance sample group is compared. It is determined that the overall survival of the high-abundance sample group is longer than that of the low-abundance sample group. The prognostic correlation verification results are output, showing that the clinical abundance parameters of the target cell subpopulation are positively correlated with the survival prognosis of the glioma patients corresponding to the glioma patient samples.

[0007] Preferably, obtaining the target subpopulation characteristic gene set and the total microglia characteristic gene set includes: Immunocured mice and control mice were constructed. Single-cell RNA was extracted from brain tumor tissue samples of the immune-cured mice and brain tumor tissue samples of the control group mice, respectively. Single-cell transcriptome sequencing was performed on the single-cell RNA to obtain single-cell transcriptome sequencing data; The Csmd3 gene was obtained from the total microglia population as a target identification gene, and the expression level of the target identification gene within the microglia population was evaluated. Based on the expression level, specific cell clusters expressing the targeted identification gene are isolated from the microglia population, and the specific cell clusters are identified as the target cell subpopulation. The characteristic genes of the target cell subpopulation are extracted, and the characteristic gene set of the target subpopulation is constructed.

[0008] Preferably, the extraction of microglia from the single-cell transcriptome sequencing data includes: Perform quality control on the single-cell transcriptome sequencing data, remove cells whose number of expressed genes is outside the preset range and whose mitochondrial gene ratio is greater than the preset threshold, and output the quality-controlled single-cell sequencing dataset. Configure a dual-cell prediction algorithm, and use the dual-cell prediction algorithm to perform dual-cell identification processing on the quality control single-cell sequencing dataset, remove cell data with dual-cell scores greater than the threshold, and generate a standard single-cell feature matrix. Gene variability analysis is performed on the standard single-cell feature matrix to extract highly variable feature genes, and principal component analysis is performed based on the highly variable feature genes to obtain the principal component analysis results. Configure a batch effect correction algorithm and a dimensionality reduction algorithm. Use the batch effect correction algorithm to perform inter-sample batch effect correction processing on the principal component analysis results, and use the dimensionality reduction algorithm to perform data dimensionality reduction, outputting a dimensionality-reduced integrated dataset. Cluster analysis was performed on the dimensionality-reduced and integrated dataset to obtain multiple cell clusters. Differentially expressed characteristic genes of these cell clusters were calculated. A chromosome copy number variation inference algorithm was configured, and the copy number variation event data generated by the algorithm was used to annotate and divide the multiple cell clusters into G422. TN -GBM tumor cell population and non-tumor cell population, from which the total microglia population is determined, and the characteristic genes of the total microglia population are extracted to construct the total microglia characteristic gene set.

[0009] Preferably, it also includes a clinical histological verification step: The glioma tissue sample was subjected to paraffin sectioning to obtain tissue section samples; Obtain the preset CSMD3 antibody and the preset IBA1 antibody; The tissue section samples were subjected to multiple immunofluorescence staining using the CSMD3 antibody and the IBA1 antibody to target and label the CSMD3 protein and IBA1 protein in the tissue section samples, thereby obtaining multiple immunofluorescence staining images. Configure an image analysis algorithm and use the image analysis algorithm to perform image feature analysis on the multiple immunofluorescence staining image to identify specific cell groups in the tissue section sample that simultaneously express CSMD3 protein and IBA1 protein, and define the specific cell groups as clinical homologous target cell groups; The spatial distribution characteristic parameters of the clinical homologous target cell population in the multiplex immunofluorescence staining image are calculated using the image analysis algorithm, and the infiltration density of the clinical homologous target cell population in the glioma tissue sample is calculated using the spatial distribution characteristic parameters.

[0010] Preferably, it also includes a clinical single-cell level validation step: Obtain the Gene Expression Omnibus (GEO) database and extract human glioma single-cell transcriptome data from the GEO database; The human glioma single-cell transcriptome data were subjected to quality control and dimensionality reduction clustering to construct a clinical single-cell sequencing dataset; Obtain preset microglia markers, and extract clinical microglia populations from the clinical single-cell sequencing dataset based on the expression characteristics of the microglia markers; The expression intensity of the clinical microglia population was evaluated using the target identification gene of the target cell subpopulation. Single cell clusters expressing the target identification gene were isolated from the clinical microglia population and defined as clinical homologous single cell subpopulations. The proportion of the clinical homologous single-cell subpopulation in the clinical microglia group is calculated. Based on the cell proportion, the presence status of the target cell subpopulation in the human clinical glioma microenvironment and its proportion in different glioma types are confirmed.

[0011] Preferably, the method further includes a step of verifying the in vitro functional intervention of the target cell subpopulation: A pre-defined mouse BV2 microglia cell line was obtained as the target cell for in vitro intervention. The promoter sequence of the mouse Csmd3 gene was obtained, and a single guide RNA sequence was designed and synthesized based on the promoter sequence of the mouse Csmd3 gene. A preset CRISPR activation system is obtained, which includes a first lentiviral particle expressing an inactivated Cas9 protein, a second lentiviral particle expressing a transcription activation domain, and a single-guide RNA lentiviral backbone plasmid. The single-guide RNA sequence is cloned into the single-guide RNA lentiviral backbone plasmid to construct a single-guide RNA expression vector. A pre-defined viral packaging helper plasmid was obtained, and a third lentiviral particle carrying the single guide RNA sequence was prepared using the single guide RNA expression vector and the viral packaging helper plasmid. As a control, a pre-defined empty lentiviral particle containing a non-targeting control sequence was used instead of the third lentiviral particle. The mouse BV2 microglia were infected with the first lentiviral particles and the first resistant drug was used for initial screening to obtain a first resistant cell population expressing inactivated Cas9 protein. Then, the first resistant cell population was infected with the second lentiviral particles, and a second resistant drug was used for secondary screening to obtain a second resistant cell population that simultaneously expresses inactivated Cas9 protein and transcription activation domain. Then, the second resistant cell population is infected with the third lentiviral particles, and a final screening is performed using the third resistant drug to obtain a stable infected cell population. The first resistance drug is blast fungicide, the second resistance drug is hygromycin B, and the third resistance drug is puromycin; Transcriptional expression data and Western blot quantitative data were obtained for the stable infected cell population. Based on the comprehensive verification results of the transcriptional expression data and the Western blot quantitative data, the stable infected overexpressing cell population was identified as MG. OE-Csmd3, BV2 The stable infection control cell population was identified as MG. NC, BV2 .

[0012] Preferably, in determining the MG OE-Csmd3,BV2 This is followed by an in vitro tumor growth inhibition assessment step: Obtain mouse GL261 glioblastoma (GL261-GBM) cells labeled with luciferase; Construct an in vitro direct cell co-culture system, and place the MG OE-Csmd3,BV2 The cells were directly mixed with the luciferase-labeled mouse GL261-GBM cells and incubated in the direct cell co-culture system for the first experimental group. Obtain a preset fluorescein substrate reagent, add the fluorescein substrate reagent to the direct cell co-culture system after incubation treatment in the first experimental group, configure a bioluminescence imaging component, use the bioluminescence imaging component to acquire bioluminescence images of the direct cell co-culture system, and extract the bioluminescence radiance value of the region of interest. An in vitro indirect cell co-culture system was constructed, and the MG cells were cultured in a pre-designed co-culture separation chamber. OE-Csmd3,BV2 The cells were physically isolated from the mouse GL261-GBM cells labeled with luciferase and seeded into the upper and lower chambers of the indirect cell co-culture system for incubation in the second experimental group. A cell viability detection component was configured, and mouse GL261-GBM cells were extracted from the lower chamber after the incubation treatment of the second experimental group. The relative cell viability data of the mouse GL261-GBM cells after indirect co-culture were measured using the cell viability detection component. Based on the bioluminescence radiance value and the relative cell viability data, the MG is determined comprehensively. OE -Csmd3,BV2 It has the ability to inhibit the proliferation of the mouse GL261-GBM cells.

[0013] Preferably, in determining the MG OE-Csmd3,BV2 This is followed by an in vivo tumor growth inhibition assessment step: First GBM cells and second GBM cells were obtained. The first GBM cells were mouse GL261-GBM cells labeled with luciferase, and the second GBM cells were mouse G422 cells. TN -GBM cells; Obtain a preset cell resuspension, and then separately add the MG... OE-Csmd3,BV2 The first co-implanted cell suspension and the second co-implanted cell suspension were mixed with the first GBM cells and the second GBM cells in the cell resuspension to construct the first co-implanted cell suspension and the second co-implanted cell suspension, respectively. Experimental mice that are immune to the first GBM cells and the second GBM cells were obtained and divided into the first tumor model experimental mouse group, the first tumor model control mouse group, the second tumor model experimental mouse group, and the second tumor model control mouse group, respectively. A stereotactic injection device was configured to inject the first co-implanted cell suspension into the brain tissue region (right striatum) of the first tumor model mouse group to construct a GL261-GBM co-implanted tumor model. The second co-implanted cell suspension was then injected into the brain tissue region (right striatum) of the second tumor model mouse group to construct a G422 co-implanted tumor model. TN -GBM co-implanted tumor model; Configure in vivo imaging detection components and tissue section staining components; For the GL261-GBM co-implanted tumor model, the in vivo imaging detection component is used to acquire in vivo image data of the GL261-GBM co-implanted tumor model to obtain tumor in vivo growth data, and a portion of the brain tissue of the GL261-GBM co-implanted tumor model is extracted for hematoxylin-eosin staining. The area of ​​the largest cross section of the tumor is measured by the tissue section staining component to obtain the first tumor burden data. Regarding the G422 TN -GBM co-implanted tumor model, its brain tissue was directly extracted and hematoxylin-eosin stained, and the area of ​​the largest cross section of the tumor was measured by the tissue section staining component to obtain the second tumor burden data; The survival time of mice from various co-implanted tumor models was integrated to obtain mouse survival data. Using the in vivo tumor growth data, the first tumor burden data, the second tumor burden data, and the mouse survival data, the MG was comprehensively determined. OE-Csmd3,BV2 It has an inhibitory effect on tumor growth in the in vivo environment.

[0014] Preferably, it also includes tumor microenvironment remodeling and tertiary lymphoid structure induction analysis steps: Brain tumor tissue samples were extracted from the experimental mouse group and processed into paraffin sections to obtain tumor tissue sections. Obtain preset CD3 antibody and preset CD20 antibody, and use the CD3 antibody and CD20 antibody to perform immunohistochemical staining on the tumor tissue section to target and label the CD3 protein and CD20 protein in the tumor tissue section, and obtain the tumor section staining image. A microscopic image analysis algorithm is configured, and the image feature analysis of the stained tumor slice image is performed using the microscopic image analysis algorithm to identify T cells expressing CD3 protein and B cells expressing CD20 protein respectively. The cell infiltration density of the T cells and the B cells in the non-tertiary lymphoid structure region within the tumor is calculated to generate immune cell infiltration abundance data. The spatial structure of T cells expressing CD3 protein and B cells expressing CD20 protein aggregates is identified using the microscopic image analysis algorithm, and the spatial structure containing T cell and B cell aggregates is defined as a tertiary lymphoid structure. The number and area of ​​the tertiary lymphoid structures in the tumor tissue section are calculated using the aforementioned microscopic image analysis algorithm, generating structural distribution density data; It also includes an immune cell memory phenotype assessment step: Single-cell transcriptome sequencing data were obtained from immune-cured mice and control mice, and the target cell subpopulations, T cell populations and B cell populations were extracted from them. Expression analysis of key regulatory factors of innate immune memory was performed to determine that key transcription factors of innate immune memory were upregulated in the target cell subpopulation within the immune-cured mice. Based on the enrichment analysis results of the differential gene set, the significantly upregulated feature score data, and the key transcription factors of the innate immune memory, it was determined that the cellular state attribute of the target cell subpopulation in the immune-cured mice was an immune-activated state and that it had an innate immune memory phenotype. Obtain multiple memory taxonomic marker genes in T cells from a predefined set, including Cd69, Itgae, Cd44, Sell, Ccr7, Il7r, and Il15ra; The expression levels of specific T cell memory group marker genes within the T cell population were calculated. By comparing the expression levels of the specific T cell memory group marker genes in the immunized cured mice and the control mice, it was determined that the T cell population in the immunized cured mice possessed tissue residency memory phenotype characteristics, and was characterized as a tissue residency memory T cell population. Obtain preset memory B cell markers, including Cxcr5 and Cr2 genes, and analyze the expression of the memory B cell markers within the B cell population; By comparing the expression of Cxcr5 and Cr2 genes in the immune-cured mice and the control mice, it was determined that the B cell population in the immune-cured mice possesses immune memory phenotype characteristics, exhibiting memory B cell population characteristics. It also includes steps for assessing tumor-killing activation status and analyzing cell communication networks in the anti-GBM immune memory microenvironment: The single-cell transcription feature matrix is ​​extracted and input into a preset cell communication analysis algorithm for calculation and processing to obtain receptor-ligand interaction data; Based on the receptor-ligand interaction data, by comparing the immune-cured mice with the control group mice, it was determined that the overall number of intercellular interactions was increased and the tumor cell self-interactions were weakened in the immune-cured mice. The enhanced communication strength between the tissue-resident memory T cells and the tumor cells was determined, and the signaling pathways and specific receptor-ligand pairs exhibiting enhanced characteristics were extracted. The signaling pathways included PARs, TRAIL and LCK signaling pathways, and the specific receptor-ligand pairs included Cd6-Alcam, Cd226-Pvr and Tnfsf10-Tnfrsf10b. The T cell population is further subdivided into multiple T cell subpopulations, including CD4.+ Tissue-resident memory T cell subsets and CD8 + Organ-resident memory T cell subsets; CD8 was identified through cell communication network analysis. + Organ-resident memory T cell subsets act as primary signal transducers and receivers, as well as CD4+. + Organ-resident memory T cell subsets act as regulators of CD8 + The main signaling agents for the activity of tissue-resident memory T cell subsets; Based on the combined analysis of the receptor-ligand interaction data, the signaling pathways, the specific receptor-ligand pairs, and the cell communication network of the T cell subpopulations, it was determined that the anti-GBM immune memory microenvironment in the immunized mice was in a state of tumor killing activation. It also includes steps for triggering an assessment of the immunotherapy response: Obtain another batch of mice G422 TN - GBM cell immune background matched experimental mice were divided into a combination therapy mouse group and a basic therapy mouse group; Using the aforementioned stereotactic injection device, the MG (metabolite) is injected. OE-Csmd3,BV2 With the mouse G422 TN - GBM cell co-implantation cell suspension was injected into the brain tissue region (right striatum) of the combined treatment mouse group and the basic treatment mouse group, respectively. Obtain a preset PD-1 immune checkpoint blocking antibody and an isotype control antibody. Inject the PD-1 immune checkpoint blocking antibody into each mouse in the combined treatment mouse group via intraperitoneal injection. Inject the isotype control antibody into each mouse in the basic treatment mouse group via intraperitoneal injection. Survival time was continuously recorded to obtain combined treatment survival data. The survival time of the combined treatment mouse group and the basic treatment mouse group was compared to determine the MG. OE-Csmd3,BV2 It has the function of inducing immunotherapy response (triggering PD-1 immune checkpoint blocking antibody response) and prolonging the survival of tumor-bearing mice; It also includes an anti-GBM immune memory assessment step: Obtain and accept the MG OE-Csmd3,BV2 With the mouse G422 TN - A group of mice that survived for a long time after co-implantation of GBM cells and reached the first preset time threshold, and age- and sex-matched normal mice were obtained as a re-challenge control group. The mouse G422 TN -GBM cells were resuspended separately in the cell resuspension to construct a secondary implantation cell suspension; Using the stereotactic injection device, the secondary implanted cell suspension was injected into the brain tissue region (left striatum) of each mouse in the long-term surviving mouse group and the re-challenge control mouse group to perform tumor secondary implantation treatment without any additional anti-tumor treatment. Using the aforementioned live imaging detection component, live imaging of the brain tissue region (left striatum) after the secondary tumor implantation treatment is performed to obtain secondary tumor growth data; The survival time after secondary implantation was continuously recorded, and the long-term surviving mouse group whose survival time reached the second preset time threshold was defined as the immune-cured mouse group. Based on the secondary tumor growth data and the long-term survival rate of tumor-bearing mice in the immune-cured mouse group, the MG was determined. OE-Csmd3,BV2 It has the function of establishing long-lasting anti-GBM immune memory.

[0015] Preferably, the application further includes: the application of the prognostic correlation verification results in the preparation of an auxiliary assessment tool for screening patients with gliomas who are sensitive to immunotherapy.

[0016] This invention provides a Csmd3 for assessing the prognosis of glioma patients. + A method for detecting microglia. It offers the following benefits: 1. This invention acquires clinical transcriptome data and glioma tissue samples, and obtains the characteristic gene sets of the target subpopulation and the total microglia characteristic gene set. A single-sample gene set enrichment analysis algorithm is configured to calculate the subpopulation enrichment score and the total microglia enrichment score, respectively. Based on the ratio of these two scores, the clinical abundance parameter of the target cell subpopulation in each glioma patient sample is determined. A survival analysis algorithm is further configured to divide patient samples into high-abundance and low-abundance groups. Combined with clinical survival data, survival prognostic assessment data is obtained, thus accurately outputting the prognostic correlation verification results showing a positive correlation between the clinical abundance parameter of the target cell subpopulation and the survival prognosis of glioma patients. The ratio-based assessment method effectively eliminates the interference of differences in the initial total microglia infiltration count between individuals, achieving accurate prognostic assessment of glioma patients and providing reliable data support for clinical treatment.

[0017] 2. This invention constructs immune-cured mice and control mice, extracts single-cell RNA and performs sequencing to obtain single-cell transcriptome sequencing data, extracts microglia from these data and obtains the Csmd3 gene as a target identification gene, assesses expression levels and isolates specific cell clusters as target cell subpopulations, thereby extracting characteristic genes to construct a target subpopulation characteristic gene set. This process effectively overcomes the limitation of conventional static models in simulating long-term immune memory states. By accurately comparing the cellular omics differences between the immune-cured microenvironment and the control group, the blindness of conventional target screening is eliminated, ensuring the accuracy and reliability of target cell population identification and laying a solid foundation for subsequent prognostic assessment.

[0018] 3. This invention obtains a pre-defined mouse BV2 microglia cell line as the target cell for in vitro intervention. Utilizing a triviral CRISPR activation system containing an expression of inactivated Cas9 protein, a transcriptional activation domain, and a single-guide RNA sequence, the mouse BV2 microglia cell line is treated with a progressive infection strategy. Multiple corresponding resistance drugs are used for stepwise screening to rigorously obtain a stable transfected cell population. Finally, transcriptional expression data and Western blot quantitative data are used to verify the generation of MG (Millage-Induced Glycerin) cells. OE-Csmd3,BV2 The study evaluated its ability to inhibit glioma cell proliferation in vitro and constructed a co-implanted tumor model to assess its in vivo tumor growth inhibition effect. Finally, combined treatment survival data were obtained to confirm its function in inducing immunotherapy response and establishing long-term anti-GBM immune memory. This comprehensively and deeply verified the key mechanism of the target cell subset in the anti-tumor process, providing a clear intervention target for the development of novel glioma immunotherapy. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the ICu screening paradigm of the present invention; Figure 2 This is a schematic diagram illustrating the quality control and pre-analysis of ICu and control mouse scRNA-seq data in this invention; Figure 3 This is a schematic diagram of the integrated analysis of ICu and control mouse scRNA-seq data from the present invention; Figure 4 This is a schematic diagram of the bulk RNA-seq data analysis of ICu and control mice in this invention; Figure 5 Csmd3 of the present invention + Confirmation of microglia and Csmd3 levels in ICu mice compared to control mice + Schematic diagram of changes in microglia; Figure 6 Csmd3 in ICu mice of the present invention +A schematic diagram illustrating the memory phenotype exhibited by microglia, T cells, and B cells; Figure 7 This is a schematic diagram of cell communication analysis of ICu and control mice integrating scRNA-seq data according to the present invention; Figure 8 This diagram illustrates the T cell subpopulation segmentation, pseudo-temporal analysis, and cell communication analysis in the integrated scRNA-seq data of ICu and control mice according to the present invention. Figure 9 This is a schematic diagram illustrating the communication analysis between the cells that make up the AGIM unit of the present invention; Figure 10 Csmd3 of the present invention + A schematic diagram illustrating the prognostic analysis of microglia subsets in the TCGA and CGGA cohorts; Figure 11 Csmd3 in clinical LGG and GBM patients of the present invention + Schematic diagram of microglia; Figure 12 Csmd3 from the public scRNA-seq data of clinical glioma patients of this invention + Schematic diagram of microglia; Figure 13 To demonstrate the effectiveness of the combined scRNA-seq and bulk RNA-seq data analysis of Csmd3 in this invention + Schematic diagram of microglia exhibiting a pro-inflammatory and anti-tumor phenotype; Figure 14 This is a schematic diagram illustrating the changes in pro-inflammatory genes after overexpression of the BV2 microglial cell line Csmd3 in the CRISPRa system of the present invention. Figure 15 The MG of the present invention OE-Csmd3,BV2 Schematic diagram of the in vitro co-culture experiment of cells and GL261-GBM cells; Figure 16 The MG of the present invention OE-Csmd3,BV2 Cells and GL261-GBM cells and G422 TN -Schematic diagram of in vivo co-implantation experiment of GBM cells; Figure 17 To verify Csmd3 using various methods of the present invention + A schematic diagram illustrating how microglia can induce the formation of TLS in GBM and trigger the responsiveness of PD-1 blocking monoclonal antibodies; Figure 18 Csmd3 of the present invention + A schematic diagram illustrating the 100% ICu rate in microglia-induced cured mice. Detailed Implementation

[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Example

[0021] See attached document Figure 1 The workflow of the immune cure screening paradigm provided by this invention may include: Primary cells were obtained and maintained using an alternating subcutaneous-intracranial in vivo passage method for primary G422 cells. TN - The biological characteristics of GBM cells were analyzed, and single-cell suspensions were prepared. Experimental mice of a predetermined strain (e.g., the Kunming mouse strain) were obtained. The right striatal region of the experimental mice was selected as the target inoculation site. Single-cell suspensions obtained through in vivo passage were inoculated into the right striatal region at a seeding rate of 5 × 10^4 cells using a stereotaxic instrument to construct G422 cells. TN -GBM orthotopic tumor-bearing mouse model (e.g.) Figure 1 G422 TN -GBM vaccination illustration shown).

[0022] Obtain a pre-selected anti-tumor drug (e.g., temozolomide) for G422. TN -Anti-tumor drug treatment intervention was implemented in the GBM orthotopic tumor-bearing mouse model, and G422 was recorded. TN -The survival time of the GBM orthotopic tumor-bearing mouse model was determined by selecting mice that survived for more than 100 days after drug treatment intervention, and these mice were identified as long-term survivors (e.g., Figure 1 (As shown in the LTS mice), where long-term survival (LTS) refers to the first tumor rechallenge experiment in long-term survival mice. The left striatum region of the long-term survival mice was selected as the target inoculation site for the first tumor rechallenge experiment, and 5 × 10^4 primary G422 cells were injected. TN GBM cells were seeded into the left striatal region (e.g.) Figure 1 (As shown in the 1°-rechallenge phase), the survival time of long-term surviving mice after the first tumor rechallenge experiment was recorded. Mice that did not receive additional drug treatment and survived for more than 100 days again were selected as immune-cured mice (e.g., ...). Figure 1 As shown in the ICu mice, ICu (Immune-cure) forms an anti-GBM immune memory microenvironment in the immune-cure mice, which is called anti-GBM immune memory (AGIM).

[0023] A second tumor rechallenge experiment was conducted on immune-cured mice (e.g.) Figure 1 (as shown in the 2°-rechallenge stage), 5×10^4 primary G422 TN GBM cells were inoculated into the right striatum region of immune-cured mice.

[0024] Brain tumor tissue samples were obtained from immune-cured mice that underwent a second tumor re-challenge experiment, and brain tumor tissue samples were also obtained from pre-constructed control mice. The control mice were tumor-bearing mice that were inoculated with the same number of primary tumor cells but did not receive drug treatment. Single-cell RNA and total RNA were extracted from the brain tumor tissue samples of immune-cured mice and control mice, respectively. Single-cell transcriptome sequencing was performed on the single-cell RNA to obtain single-cell transcriptome sequencing data, and total RNA was sequenced to obtain transcriptome sequencing data. The single-cell transcriptome sequencing data and the transcriptome sequencing data together constituted multi-omics sequencing data.

[0025] Bioinformatics integration analysis was performed on multi-omics sequencing data. Quality control, dimensionality reduction, clustering, and cell subpopulation annotation were conducted on single-cell transcriptome sequencing data from the multi-omics sequencing data. Cell composition differences between immunized cured mice and control mice in the anti-GBM immune memory microenvironment were compared. Combined with enrichment analysis results from the transcriptome sequencing data in the multi-omics sequencing data, a target cell subpopulation with increased numbers was identified from the anti-GBM immune memory microenvironment. The target cell subpopulation was identified as Csmd3. + For microglia, transcriptomic characteristic parameters of the target cell subpopulation were further extracted based on multi-omics sequencing data.

[0026] After extracting transcriptomic characteristic parameters of target cell subpopulations based on multi-omics sequencing data, the present invention provides an overall system architecture that may include: an animal model discovery module, a clinical validation module, a functional intervention module, an in vivo assessment module, and a mechanism validation module.

[0027] The animal model discovery module is used to receive multi-omics sequencing data extracted from brain tissue samples of immune-cured mice, and to identify target cell subpopulations and generate their transcriptomics characteristic parameters based on the multi-omics sequencing data.

[0028] The clinical validation module is used to receive transcriptomics feature parameters, obtain clinical transcriptomics data and analysis results of glioma tissue samples. The clinical transcriptomics data is bulk RNA-seq data, and outputs the prognostic correlation validation results of the abundance parameters of the target cell subpopulation with the survival prognosis of glioma patients.

[0029] The functional intervention module is used to acquire data generated from functional intervention experiments based on the mouse BV2 microglia cell line and mouse GL261-GBM. The experiments utilize the CRISPR activation system to generate MG cells. OE-Csmd3,BV2 and MG OE-Csmd3,BV2 We co-cultured tumor cells derived from mouse GL261-GBM in vitro to generate in vitro tumor suppression data.

[0030] The liveness assessment module is used to obtain MG-based... OE-Csmd3,BV2 Compared with mouse GL261-GBM and mouse G422 TN The co-implanted tumor model data was constructed using GBM glioma cell lines. The in vivo assessment module evaluated the co-implanted tumor model, obtained tumor growth volume data and mouse survival data, and received histological measurement results of brain tumor tissue samples, extracting structural distribution density data.

[0031] The in vivo assessment module is used to evaluate immunotherapy response and long-term immune memory function in stages. The in vivo assessment module receives co-implanted models (including MG) for the initial vaccination. OE-Csmd3,BV2 Combined therapy survival data (including co-implantation with pre-defined type GBM cells) following injection of a pre-defined PD-1 immune checkpoint blocking antibody were used to determine MG OE-Csmd3,BV2 The triggering characteristics of drug sensitivity to immunotherapy are then used. The in vivo assessment module receives secondary tumor growth and survival data after tumor reimplantation (i.e., re-entry of homologous GBM cells without any treatment) in mice that have achieved long-term survival. Based on the data, it is confirmed whether long-term anti-GBM immune memory has been established in the mice (i.e., immune cure has been achieved).

[0032] The mechanism validation module utilizes data from extracted microglia total RNA, tumor tissue total RNA, tissue section microscopic analysis results, and multiplex in vivo detection components (including in vivo imaging detection components) to perform multi-level molecular, tissue, and in vivo MG analysis. OE-Csmd3,BV2 The mechanism for promoting tumor microenvironment remodeling and establishing long-term anti-GBM immune cure was comprehensively validated.

[0033] After clarifying the overall system architecture and the division of labor among the modules, the specific implementation steps of this invention first involve the animal model discovery module performing the screening and confirmation of target cells, as detailed in the appendix. Figure 2 The present invention provides a method for integrating single-cell transcriptome sequencing data and identifying target cells, which may include: Brain tumor tissue samples were extracted from immune-cured mice and control mice for sequencing to generate multi-omics sequencing data. The animal model discovery module acquired and separated this multi-omics sequencing data to obtain single-cell transcriptome sequencing data. The animal model discovery module then performed quality control on the single-cell transcriptome sequencing data, removing cells with gene expression counts between 200 and 10,000 and cells with a mitochondrial gene proportion greater than 15%. The quality-controlled single-cell sequencing dataset (e.g., ...) was then output. Figure 2 A, Figure 2 As shown in B), the animal model discovery module uses a two-cell prediction algorithm to perform two-cell identification processing on the quality-controlled single-cell sequencing dataset. The two-cell prediction algorithm is used to calculate the two-cell score parameters of each cell in the quality-controlled single-cell sequencing dataset (e.g., ...). Figure 2 As shown in C), the animal model discovery module removes cell data with a two-cell score greater than 0.15 based on the two-cell score parameter, and generates a standard single-cell feature matrix.

[0034] See attached document Figure 2 The animal model discovery module performs data normalization and data scaling on the standard single-cell feature matrix.

[0035] The animal model detection module performs gene variability analysis on the scaled standard single-cell feature matrix, identifying multiple highly variable feature genes. It then extracts the top 2000 highly variable feature genes and performs principal component analysis (PCA) based on these genes. The module obtains the PCA results and applies batch effect correction algorithms to the PCA results (e.g., ...). Figure 2 As shown in D), the animal model discovery module uses a dimensionality reduction algorithm to perform dimensionality reduction on the data based on the first 15 principal components after correction, and outputs a dimensionality-reduced integrated dataset.

[0036] The animal model discovery module performs cluster analysis on the dimensionality-reduced integrated dataset at a resolution of 0.8, constructs a shared nearest neighbor network graph, and obtains multiple cell clusters generated by the cluster analysis (such as...). Figure 2 As shown in E), the animal model discovery module calculates differentially expressed characteristic genes from multiple cell clusters. The animal model discovery module acquires copy number variation event data generated by the chromosome copy number variation inference algorithm and combines it with differentially expressed characteristic genes from multiple cell clusters (such as...). Figure 2 F, Figure 2 As shown in G), multiple cell cluster annotations are divided into G422. TN -GBM tumor cell groups and non-tumor cell groups.

[0037] The animal model discovery module classifies non-tumor cell groups into cell subpopulations, such as... Figure 3As shown in Figure A, the animal model discovery module uses a UMAP plot to display all cell types in the dimensionality-reduced integrated dataset (i.e., region A); as... Figure 3 As shown in B, the animal model discovery module uses a heatmap to display the expression of marker genes for each cell type (i.e., region B).

[0038] The animal model discovery module identifies the total microglia population from non-tumor cell populations and extracts characteristic genes from this population to construct a characteristic gene set. The module calculates differentially expressed genes among different subpopulations within the total microglia population relative to other cell populations in the non-tumor cell population. The module obtains the Csmd3 gene as a target identification gene and uses differentially expressed genes to assess the expression level of the target identification gene within the total microglia population. Based on the expression level, the module isolates specific cell clusters expressing the target identification gene from the total microglia population. These specific cell clusters are then identified as the target cell subpopulation, and the target cell subpopulation is identified as Csmd3. + Microglia, animal model discovery module statistically reduced and integrated the cell composition ratio of immune-cured mice and control mice in the dataset (e.g. Figure 3 As shown in C), the animal model discovery module determined that the proportion of target cell subsets in immune-cured mice was higher than that in control mice.

[0039] The animal model discovery module identifies T cell and B cell populations from non-tumor cell populations. The animal model discovery module confirms that the distribution abundance of T cell and B cell populations in immune-cured mice is higher than that in control mice. The target cell subsets, T cell populations, and B cell populations together constitute the anti-GBM immune memory microenvironment of immune-cured mice.

[0040] After the target cell subsets, T cell populations, and B cell populations jointly constitute the anti-GBM immune memory microenvironment in immunocured mice, refer to the appendix. Figure 4 The present invention provides a method for dual verification of transcriptome data and histology, which may include: The animal model discovery module separates and obtains transcriptome sequencing data from multi-omics sequencing data. The animal model discovery module is equipped with a single-sample gene set enrichment analysis algorithm: specifically, the single-sample gene set enrichment analysis algorithm refers to ssGSEA. Based on single-cell transcriptome sequencing data, the animal model discovery module extracts characteristic genes of target cell subpopulations and constructs a characteristic gene set of the target subpopulation. At the same time, based on the microglia group in the single-cell transcriptome sequencing data, the animal model discovery module extracts microglia characteristic genes and constructs a total microglia characteristic gene set.

[0041] The animal model discovery module uses a single-sample gene set enrichment analysis algorithm to process transcriptome sequencing data. The module inputs the target subpopulation characteristic gene set into the algorithm to calculate the subpopulation enrichment score (e.g., ...) for each sample in the transcriptome sequencing data. Figure 4 As shown in A), the animal model discovery module inputs the total microglia characteristic gene set into the single-sample gene set enrichment analysis algorithm to calculate the total microglia enrichment score of each sample in the transcriptome sequencing data.

[0042] The animal model detection module calculates the ratio between the subpopulation enrichment score and the total microglia enrichment score to obtain the relative proportion of the target cell subpopulation in the total microglia (i.e., the relative proportion parameter, such as...). Figure 4 As shown in B).

[0043] The animal model discovery module determined that the enrichment degree and relative proportion of target cell subpopulations in immunized cured mice were higher than those in control mice by comparing the subpopulation enrichment scores and relative proportion parameters between immunized cured mice and control mice. The enrichment analysis results of transcriptome sequencing data verified the existence of target cell subpopulations in the anti-GBM immune memory microenvironment of immunized cured mice.

[0044] See attached document Figure 5 Normal mice were obtained and their brain tissue was extracted. The brain tissue was sectioned to obtain brain tissue slice samples. Pre-set CSMD3 antibody and pre-set IBA1 antibody were obtained. The brain tissue slice samples were then subjected to immunofluorescence double staining using the CSMD3 antibody and IBA1 antibody to target and label the CSMD3 protein and IBA1 protein in the brain tissue slice samples, and the immunofluorescence staining pattern was obtained (e.g., ...). Figure 5 (As shown in A).

[0045] The animal model discovery module obtained and identified cells expressing both CSMD3 and IBA1 proteins in the cortical region of brain tissue slices using immunofluorescence staining. The immunofluorescence staining confirmed the presence of the target cell subset in the brain tissue. The animal model discovery module also obtained mouse BV2 microglia and human HMO6 microglia. The animal model discovery module jointly identified the mouse BV2 microglia and human HMO6 microglia as the validation cell lines.

[0046] Total protein was extracted from the validation cell line and detected using Western blotting technology to obtain Western blotting data (e.g., ...). Figure 5As shown in B), the animal model discovery module uses Western blot data to determine the expression of CSMD3 protein in the validation cell line. It then uses CSMD3 antibody to perform immunofluorescence staining on the validation cell line, acquiring immunofluorescence staining images. The animal model discovery module combines these staining images to verify the intracellular localization of CSMD3 protein (e.g., ...). Figure 5 C Figure 5 As shown in D), both Western blot data and immunofluorescence staining results demonstrated that CSMD3 protein is endogenously expressed in both mouse and human microglia. Brain tissue sections were obtained from immunized cured mice (original tumor inoculation area) and control mice. Immunofluorescence double staining was performed on the brain tissue sections using CSMD3 antibody and IBA1 antibody to obtain tissue staining images (e.g., as shown in D). Figure 5 E is shown.

[0047] Animal model detection module compared staining results and determined that the number of target cell subpopulations in brain tumor tissue sections of immune-cured mice was higher than that in brain tumor tissue sections of control mice. The staining results were consistent with the analysis results of single-cell transcriptome sequencing data, further confirming the increase of target cell subpopulations in immune-cured mice.

[0048] After the analysis of staining results and single-cell transcriptome sequencing data was consistent, further confirming the increase of the target cell subset in immune-cured mice, refer to the appendix... Figure 6 The present invention provides a method for analyzing the phenotype and state of immune cells, which may include: The animal model discovery module extracts target cell subpopulations from single-cell transcriptome sequencing data, originating from immunized-cured mice and control mice, respectively. The module then calculates differentially expressed genes (DEGs) in the target cell subpopulations of immunized-cured mice relative to those in control mice, obtaining the set of upregulated genes.

[0049] The animal model discovery module is equipped with a biological pathway enrichment analysis algorithm (such as GO enrichment analysis). This module uses the algorithm to process the upregulated gene set and obtain biological pathway enrichment data (such as...). Figure 6 As shown in A), the animal model detection module simultaneously calculates the characteristic scores of specific inflammatory signaling pathways within the target cell subpopulation, generating phenotypic quantitative data (such as...). Figure 6 As shown in B).

[0050] Biological pathway enrichment data included characteristic parameters of immune inflammatory response pathways such as tumor necrosis factor (TNF) signaling pathway activation and leukocyte migration. Phenotypic quantitative data included signaling characteristic scores of IL-1, TNF, IL-6, and NF-κB. Furthermore, the animal model discovery module identified Csmd3 through expression analysis of key regulatory factors of innate immune memory (IIM).+ Microglia upregulate key transcription factors of innate immune memory (IIM), such as... Figure 6 (as shown in C) The animal model discovery module used enrichment analysis of differentially expressed gene sets to significantly upregulate characteristic score data and key transcription factors of innate immune memory (IIM), and determined that the cellular state of the target cell subset in immune-cured mice was in an immune-activated state, exhibiting the innate immune memory (IIM) phenotype.

[0051] The animal model discovery module extracts T cell populations from single-cell transcriptome sequencing data and acquires preset T cell memory population marker genes (such as CD44, CD62L, etc.) to construct a T cell memory marker gene set.

[0052] The animal model discovery module calculates the expression of specific T cell memory group marker genes (including Cd69, Itgae, Cd44, Sell, Ccr7, Il7r, and Il15ra) within T cell populations (e.g., Figure 6 As shown in D), the animal model discovery module determined that the T cell population in the immunized mice possessed tissue residency memory phenotype characteristics by comparing the expression levels of the aforementioned specific genes in immunized cured mice and control mice.

[0053] The animal model discovery module extracts B cell populations from single-cell transcriptome sequencing data, acquires pre-defined memory B cell markers (including Cxcr5 and Cr2 genes), and analyzes the expression of memory B cell markers within the B cell population (e.g., ...). Figure 6 E is shown.

[0054] The animal model discovery module, by comparing the expression of Cxcr5 and Cr2 genes in immunized cured mice and control mice, determined that the B cell population in immunized cured mice possesses immune memory phenotype characteristics. After confirming the memory phenotypes of the target subset, T cell population, and B cell population in immunized cured mice, the animal model discovery module, referring to the appendix... Figure 7 and Figure 8 The present invention provides a method for analyzing the tumor killing activation state in the anti-GBM immune memory microenvironment, which may include: The single-cell transcription feature matrix is ​​extracted. The animal model discovery module is equipped with a cell communication analysis algorithm. The single-cell transcription feature matrix is ​​input into the cell communication analysis algorithm for calculation and processing to obtain receptor-ligand interaction data.

[0055] The animal model discovery module used receptor-ligand interaction data to determine that in immune-cured mice, the overall number of cell-to-cell communication interactions increased, but the communication strength decreased. The decrease in strength was mainly attributed to G422. TN-GBM tumor cells exhibit weakened self-interaction, indicating suppressed tumor growth (e.g., ...). Figure 7 (As shown in AB). Conversely, the target cell subset, T RM Cells, B M Cells and G422 TN -Communication between GBM tumor cells is generally enhanced, among which T RM Cells and G422 TN The most significant increase in communication between GBM tumor cells is observed, specifically: T RM Cells include tissue-resident memory T cells, such as TRM cells, and B cells. M The cell is a memory B cell.

[0056] Furthermore, signaling pathways exhibiting enhanced characteristics were extracted, revealing that in the anti-GBM immune memory microenvironment of immunized mice, PARs (including Gzma-Pard3 and Gzma-F2r), TRAIL, and LCK signaling pathways were significantly increased (e.g., Figure 7 (As shown in CE); specific receptor-ligand pairs were extracted, and it was determined that specific signals mediated by the Cd6-Alcam, Cd226-Pvr, and Tnfsf10-Tnfrsf10b receptor-ligand pairs constituted the specific T cells in immune-cured mice. RM -G422 TN -GBM interaction networks (such as Figure 7 As shown in F).

[0057] See attached document Figure 8 AC, the animal model discovery module will T / T RM Cells are subdivided into CD4 + CD8 + Double positive (DP, CD4) + CD8 + ) and double negative (DN, CD4) - CD8 - Subgroups were identified, and pseudo-temporal analysis was used to determine their potential differentiation trajectories. The animal model discovery module, through comparison, determined that CD8+ in the anti-GBM immune memory microenvironment of immunized cured mice... + T RM Cells were significantly enriched, and CD4 + T RM The cells also proliferated and exhibited an inflammatory phenotype (CD69). hi CD103 lo )(like Figure 8 (As shown in D). Further analysis of the communication network confirmed that in the activated microenvironment, CD8... + T RM Cells act as the primary signal senders and receivers, while CD4... +T RM Cells primarily act as signal transducers to regulate CD8. + T RM activity (e.g.) Figure 8 (As shown in E). In summary, the animal model discovery module, by integrating the aforementioned receptor-ligand communication network and T cell subpopulation status, determined that the anti-GBM immune memory microenvironment in immunized healed mice was in an activated state of anti-tumor killing.

[0058] The animal model discovery module extracts data from the target cell subset (Csmd3). + Microglia), B / B M Cells and CD4 + / CD8 + T RM Anti-GBM immune memory units, composed of cells, were calculated and established to create specific cell communication regulatory networks among the constituent cells within these units (e.g., ...). Figure 9 (As shown).

[0059] The animal model discovery module establishes a cell communication regulatory network for target cell subsets, T cell populations, and B cell populations within the anti-GBM immune memory microenvironment based on signaling pathway connectivity strength.

[0060] After establishing the cell communication regulatory network of target cell subsets, T cell populations, and B cell populations in the anti-GBM immune memory microenvironment based on signal pathway connectivity strength in the animal model discovery module, refer to the appendix. Figure 10 The present invention provides a survival prognostic analysis method based on a public database, which may include: The clinical validation module acquires data from the Cancer Genome Atlas (TCGA) database and the Chinese Glioma Genome Atlas (CGGA) database, and uses these two databases as the data sources for clinical validation. The clinical validation module obtains TCGA glioma transcriptome data and TCGA survival data from the TCGA database, and CGGA glioma transcriptome data and CGGA survival data from the CGGA database.

[0061] The clinical validation module is equipped with a batch effect correction algorithm. The clinical validation module uses the batch effect correction algorithm to eliminate the systematic error between TCGA glioma transcriptome data and CGGA glioma transcriptome data, integrates the corrected data to generate clinical transcriptome data, and integrates TCGA survival data with CGGA survival data to generate clinical survival data that matches the clinical transcriptome data.

[0062] The clinical validation module acquires the target subpopulation characteristic gene set and the total microglia characteristic gene set constructed by the animal model discovery module. The clinical validation module is equipped with a single-sample gene set enrichment analysis algorithm. The clinical validation module uses the single-sample gene set enrichment analysis algorithm to calculate the subpopulation enrichment score of each patient sample in the clinical transcriptome data relative to the target subpopulation characteristic gene set and the total microglia enrichment score relative to the total microglia characteristic gene set. The clinical validation module uses the subpopulation enrichment score and the total microglia enrichment score to calculate the ratio between the two and determine it as the relative clinical abundance parameter of the target cell subpopulation in each patient sample in the clinical transcriptome data.

[0063] The clinical validation module determines the relative clinical abundance parameter (i.e., the ratio of subset enrichment score to total microglia enrichment score) as the final clinical abundance parameter of the target cell subset in each patient sample (e.g., ...). Figure 10 A, Figure 10 As shown in C), the clinical validation module is equipped with survival analysis algorithms, including the Kaplan-Meier algorithm and the Log-rank test algorithm. The clinical validation module calculates the median value of the clinical abundance parameter and uses it as a threshold to divide the patient samples corresponding to the clinical transcriptome data into high abundance sample groups and low abundance sample groups.

[0064] The clinical validation module inputs the classification results of the high-abundance sample group, the classification results of the low-abundance sample group, and clinical survival data into the survival analysis algorithm for calculation, and obtains survival prognostic assessment data (such as...). Figure 10 B. Figure 10 (as shown in D).

[0065] The clinical validation module uses survival prognostic assessment data to compare the overall survival of the high-abundance sample group and the low-abundance sample group. The clinical validation module determines that the overall survival of the high-abundance sample group is significantly longer than that of the low-abundance sample group. The clinical validation module outputs a prognostic correlation validation result that the clinical abundance parameter of the target cell subset (i.e., the ratio of the subset enrichment score to the total microglia enrichment score) is positively correlated with the survival prognosis of glioma patients.

[0066] After the clinical validation module outputs the prognostic correlation validation results showing a positive correlation between the clinical abundance parameters of the target cell subset and the survival prognosis of glioma patients, refer to the appendix. Figure 11 Appendix Figure 12 The present invention provides a clinical histological and single-cell level verification method, which may include: Glioma tissue samples were obtained, and paraffin sections were prepared from the glioma tissue samples. Pre-defined CSMD3 and IBA1 antibodies were acquired. Multiplex immunofluorescence staining was performed on the tissue sections using the CSMD3 and IBA1 antibodies to target and label the CSMD3 and IBA1 proteins within the glioma tissue samples. Multiplex immunofluorescence staining images were obtained (e.g., ...). Figure 11 (As shown in A).

[0067] The clinical validation module is equipped with an image analysis algorithm. The clinical validation module receives multiple immunofluorescence staining images and uses the image analysis algorithm to analyze the image features of the multiple immunofluorescence staining images to identify specific cell groups in the tissue microarray samples that simultaneously express CSMD3 and IBA1 proteins. The clinical validation module defines the specific cell groups as clinical homologous target cell groups.

[0068] The clinical validation module uses image analysis algorithms to calculate the spatial distribution characteristics of clinically homologous target cell populations in multiplex immunofluorescence staining images. It then uses these spatial distribution characteristics to calculate the infiltration density of the clinically homologous target cell populations in glioma tissue samples (e.g., ...). Figure 11 B. Figure 11 As shown in C), the clinical validation module uses the analysis results of multiplex immunofluorescence staining images and the infiltration density as the histological validation results of the clinical homologous target cell population. The histological validation results of the clinical homologous target cell population confirm the existence of a cell population with homologous protein expression characteristics to the target cell subpopulation in human glioma tissue.

[0069] The clinical validation module acquires the Gene Expression Omnibus (GEO) database and extracts human glioma single-cell transcriptome data from the GEO database (e.g., Figure 12 A, Figure 12 (As shown in B) Specifically, single-cell transcriptome data refers to scRNA-seq data. The clinical validation module performs quality control and dimensionality reduction clustering on human glioma single-cell transcriptome data to construct a clinical single-cell sequencing dataset (such as...). Figure 12 C Figure 12 (as shown in D).

[0070] The clinical validation module acquires preset microglia markers, including the AIF1 gene and the GPR34 gene. The clinical validation module analyzes the expression of microglia markers in each cell in the clinical single-cell sequencing dataset and extracts the clinical microglia population.

[0071] The clinical validation module uses the CSMD3 gene as a target identification gene, evaluates the expression intensity of the target identification gene within clinical microglia, and isolates specific cell clusters expressing the CSMD3 gene from the clinical microglia population, defining them as clinical homologous single-cell subpopulations.

[0072] The clinical validation module calculates the proportion of clinically homologous single-cell subsets in patients with low-grade gliomas, newly diagnosed glioblastomas, and recurrent glioblastomas, as well as the percentage of clinically homologous single-cell subsets in the total clinical microglia population (e.g., ...). Figure 12 E, Figure 12 As shown in F), the histological verification results of the clinical homologous target cell population and the single-cell transcriptomic verification results of the clinical homologous single cell subpopulation jointly confirmed the existence of the target cell subpopulation in the human clinical glioma microenvironment, and that it decreased with the increase of glioma malignancy.

[0073] After the histological validation results of the clinically homologous target cell population and the single-cell transcriptomic validation results of the clinically homologous single-cell subset jointly confirmed the existence of the target cell subset in the human clinical glioma microenvironment and its anti-tumor immunological properties, refer to the appendix. Figure 13 The present invention provides a method for cell-targeted construction and validation, which may include: Obtain the mouse BV2 microglia cell line, obtain the mouse Csmd3 gene, obtain the promoter sequence of the mouse Csmd3 gene, and design and synthesize a single guide RNA sequence (e.g., the nucleotide sequence shown in SEQ ID NO:1) based on the promoter sequence of the mouse Csmd3 gene.

[0074] Obtaining a CRISPR activation system (such as...) Figure 13As shown in Figure A, the CRISPR activation system includes a first lentiviral particle expressing inactivated Cas9 protein, a second lentiviral particle expressing transcription activation domains (e.g., p65), and a single-guideRNA lentiviral backbone plasmid. The single-guideRNA sequence is cloned into the single-guideRNA lentiviral backbone plasmid to construct a single-guideRNA expression vector. A pre-defined viral packaging helper plasmid is obtained. Using the single-guideRNA expression vector and the viral packaging helper plasmid, a third lentiviral particle carrying the single-guideRNA sequence is prepared. The first lentiviral particle is used to infect the mouse BV2 microglia cell line, and a first-line antibiotic is used for initial screening to obtain a first-line resistant cell population expressing inactivated Cas9 protein. The second lentiviral particle is then used to infect the first-line resistant cell population, and a second-line antibiotic is used for secondary screening to obtain a second-line resistant cell population simultaneously expressing inactivated Cas9 protein and transcription activation domains. Finally, the third lentiviral particle is used to infect the second-line resistant cell population, and a third-line antibiotic is used for final screening. A stable transfected cell population was obtained. The first resistance drug was blastomycin, the second resistance drug was hygromycin B, and the third resistance drug was puromycin. As a control, empty lentiviral particles containing non-target control sequences were used to replace the third lentiviral particles. Mouse BV2 microglia were co-infected to construct a control group. Total RNA was extracted from the stable transfected cell population and the control group cells. The total microglia RNA included messenger RNA from the mouse Csmd3 gene. The total microglia RNA was reverse transcribed into complementary DNA. The complementary DNA was amplified and detected using a quantitative reverse transcription PCR detection kit to obtain transcriptional expression data (e.g., ...). Figure 13 (as shown in D).

[0075] The functional intervention module receives transcriptional expression data, confirming that the messenger RNA expression level of the mouse Csmd3 gene in the stable transfected cell population is higher than that in the control group. Total protein from microglia in both the stable transfected and control cell populations is extracted, and the total protein from microglia is detected using a Western blot assay to obtain quantitative Western blot data (e.g., ...). Figure 13 B. Figure 13 (as shown in C).

[0076] The functional intervention module received and utilized Western blot quantitative data to determine that the expression abundance of CSMD3 protein in the stable transfected cell population was higher than that in the control group. Based on the comprehensive verification results of transcriptional expression data and Western blot quantitative data, the functional intervention module identified the stable transfected cell population as MG. OE-Csmd3,BV2 Meanwhile, refer to the appendix Figure 13 This invention provides a method for verifying the expression of pro-inflammatory genes overexpressed in microglia, which may include a mechanism verification module that uses quantitative reverse transcription PCR detection data to determine the transcriptional level data of pro-inflammatory genes (e.g., Figure 13 (as shown in E), specifically: obtaining MGOE-Csmd3,BV2 In addition to control cells, total RNA was extracted, and specific amplification primers (e.g., primer sequences shown in SEQ ID NO: 2-7) targeting the interleukin 1β gene, tumor necrosis factor gene, and interleukin 6 gene were obtained. Amplification and detection were performed using quantitative reverse transcription PCR. The mechanism verification module determined MG. OE-Csmd3,BV2 The expression levels of messenger RNA of Il1b, Tnf, and Il6 in the middle cells were significantly higher than those in the control group, confirming MG. OE-Csmd3,BV2 Following the pro-inflammatory activation phenotype, refer to the appendix. Figure 14 Appendix Figure 15 This invention provides a method for evaluating cellular antitumor activity, which may include, in terms of mechanism exploration, a functional intervention module using single-cell and Bulk transcriptome sequencing data to analyze Csmd3. + Differentially expressed genes (DEGs) and related signaling pathways between microglia and other microglia subsets, ligand receptor analysis, and gene correlation analysis (e.g.) Figure 14 As shown in A-14D), Csmd3 was determined based on data analysis. + Microglia significantly upregulated major histocompatibility complex II (MHC-II) characteristic genes and phagocytosis-related genes (such as... Figure 14 As shown in A-14B), it possesses dendritic cell (DC)-like antigen-presenting characteristics (such as...). Figure 14 As shown in C), its transcriptional characteristics are strongly positively correlated with the expression of pro-inflammatory cytokines and T cell marker genes (e.g., ...). Figure 14 As shown in Figure D), the target cell subset exhibited a strong pro-inflammatory and anti-tumor phenotype at the omics level. Mouse GL261-GBM cells labeled with luciferase were obtained, and an in vitro direct cell co-culture system was constructed to culture MG cells. OE-Csmd3,BV2 Direct contact mixing with mouse GL261-GBM cells labeled with luciferase (e.g.) Figure 15 As shown in Figure A), the cells were placed in a direct cell co-culture system for the first experimental group incubation treatment. A pre-set fluorescein substrate reagent was obtained and added to the direct cell co-culture system after the first experimental group incubation treatment. A bioluminescence imaging component was configured, and bioluminescence images of the direct cell co-culture system were acquired using the bioluminescence imaging component. The bioluminescence radiance values ​​of the region of interest were extracted, and an in vitro indirect cell co-culture system (such as...) was constructed. Figure 15 As shown in Figure A), MG is separated using a pre-designed co-culture separation chamber. OE-Csmd3,BV2Mouse GL261-GBM cells labeled with luciferase were physically isolated and seeded separately in the upper and lower chambers of an indirect cell co-culture system for a second experimental group incubation treatment. A cell viability detection kit was prepared, and after the second experimental group incubation treatment, mouse GL261-GBM cells were extracted from the lower chamber. The relative cell viability data of the indirectly co-cultured mouse GL261-GBM cells were measured using the cell viability detection kit. Based on the bioluminescent radiance value and the relative cell viability data, the MG (molecular weight) was determined. OE-Csmd3,BV2 It has the ability to inhibit the proliferation of mouse GL261-GBM cells (e.g. Figure 15 (as shown in B-15E).

[0077] After confirming that the target cell subset possesses anti-tumor proliferation function in vitro, refer to the appendix. Figure 16 The present invention provides a method for constructing an in vivo co-implantation model and evaluating tumor growth inhibition, which may include: obtaining a first GBM cell and a second GBM cell, wherein the first GBM cell is a mouse GL261-GBM cell labeled with luciferase, and the second GBM cell is a mouse G422 cell. TN GBM glioblastoma cells were obtained from a pre-defined cell resuspension, and MG cells were separately... OE-Csmd3,BV2 The first and second GBM cells were mixed in a cell resuspension to construct the first and second co-implanted cell suspensions, respectively. Experimental mice with immune backgrounds matched to the first and second GBM cells were obtained and divided into four groups: a first tumor model mouse group, a first tumor model control mouse group, a second tumor model mouse group, and a second tumor model control mouse group. A stereotactic injection device was configured, and the first co-implanted cell suspension was injected into the brain tissue region of the first tumor model mouse group to construct the GL261-GBM co-implanted tumor model. The second co-implanted cell suspension was injected into the brain tissue region of the second tumor model mouse group to construct the G422 co-implanted tumor model. TN - A GBM co-implanted tumor model was developed, equipped with an in vivo imaging detection component and a tissue section staining component. For the GL261-GBM co-implanted tumor model, the in vivo imaging detection component acquired in vivo image data to obtain tumor in vivo growth data. A portion of brain tissue from the GL261-GBM co-implanted tumor model was extracted and stained with hematoxylin and eosin. The tissue section staining component measured the area of ​​the largest cross-section of the tumor to obtain the initial tumor burden data. For G422... TN- A GBM co-implanted tumor model was used, with brain tissue directly extracted and stained with hematoxylin and eosin. The area of ​​the largest cross-section of the tumor was measured using a tissue section staining component to obtain the second tumor burden data. Survival time was integrated to obtain mouse survival data for each co-implanted tumor model. Using in vivo tumor growth data, first tumor burden data, second tumor burden data, and mouse survival data, the MG (tumor leukemia) was comprehensively determined. OE-Csmd3,BV2 It has an inhibitory effect on tumor growth in the in vivo environment.

[0078] To further verify MG OE-Csmd3,BV2 For the general applicability of antitumor activity, please refer to the appendix. Figure 16 This invention provides a method for experimental verification of co-implantation of multiple gliomas in vivo. The mechanism verification module is constructed based on mouse G422. TN The detailed implementation and validation steps for the GBM cell co-implantation model are as follows: First, obtain mouse G422 TN -GBM primary cells, and obtain the preset cell resuspension, G422 TN GBM cells were not treated with luciferase labeling, and the prepared MG cells were... OE-Csmd3,BV2 With mouse G422 TN GBM cells were thoroughly mixed in a cell resuspension according to a predetermined ratio to construct a second co-implantation cell suspension. Simultaneously, as a parallel control, control group microglia were mixed with mouse G422 cells. TN GBM cells were mixed in a cell resuspension to construct a second control cell suspension. Next, cells similar to mouse G422 cells were obtained. TN - Kunming wild-type experimental mice with GBM cell immune background were randomly divided into a second experimental mouse group and a second control mouse group. A brain stereotactic injection device was configured to accurately inject the second co-implanted cell suspension into the brain tissue region of the second experimental mouse group using the device, and the second control cell suspension was injected into the brain tissue region of the second control mouse group with the same volume and coordinates.

[0079] Then, a morphological assessment of tumor growth inhibition was performed. After the preset assessment period, brain tissue from the second experimental mouse group and the second control mouse group was extracted and stained with hematoxylin and eosin. Through microscopic observation, the area of ​​the largest cross-section of the tumor in the brain tissue slices was accurately measured and compared to objectively reflect the growth burden of the solid tumor, thereby obtaining the second tumor burden data (e.g., Figure 16 As shown in F).

[0080] To continuously assess the survival benefit, the mechanism validation module continuously observes and receives survival time records from each experimental mouse in the second experimental mouse group and each control mouse in the second control mouse group, and then plots and obtains the second survival data (such as...). Figure 16 As shown in G).

[0081] The mechanism validation module used second tumor burden data and second survival data for comprehensive comparison: it determined that the tumor burden in the second experimental mouse group was significantly lower than that in the second control mouse group, and the median survival in the second experimental mouse group was significantly longer than that in the second control mouse group. The mechanism validation module is based on the above-mentioned GL261-GBM model and G422. TN - The dual in vivo validation results of the GBM model comprehensively confirm: MG OE-Csmd3,BV2 In glioma models with different genetic backgrounds and tumor microenvironments, it exhibits significant, stable, and universally applicable anti-tumor growth and prolongation of survival functions.

[0082] See attached document Figure 17 This invention provides a method for tumor microenvironment remodeling analysis and triggering immunotherapy response assessment. Based on the aforementioned findings on anti-tumor proliferation, this invention further explores the role of tumor microenvironment remodeling (MG). OE-Csmd3,BV2 The detailed implementation steps for inducing the formation of tertiary lymphoid structures and triggering the responsiveness of immune checkpoint blockers in the glioma microenvironment are as follows: In order to be on G422 TN The GBM model further validates the tumor microenvironment remodeling mechanism. The mechanism validation module performs morphological assessment of the tertiary lymphoid structure, extracts brain tumor tissue samples from mice in the second experimental mouse group and the second control mouse group, sections the brain tumor tissue samples to obtain second tumor tissue sections, and obtains preset CD20 antibody (B cell marker) and preset CD3 antibody (T cell marker). Immunohistochemical staining of the second tumor tissue sections is performed using CD20 and CD3 antibodies to target CD20 and CD3 proteins within the double-labeled sections. The mechanism validation module uses a built-in microscopic image analysis algorithm to identify the spatial structure of tightly clustered T cells and B cells (i.e., the tertiary lymphoid structure) and accurately calculates the number and area data of the tertiary lymphoid structures in the second tumor tissue sections (e.g., ...). Figure 17 D、 Figure 17 E, Figure 17 F, Figure 17 (As shown in G), the mechanism verification module used quantity and area data to determine that the number and area of ​​tertiary lymphoid structures in the second experimental mouse group were significantly higher than those in the second control mouse group, confirming the application of MG. OE-Csmd3,BV2 It can effectively induce and promote the formation of tertiary lymphoid structures in the tumor microenvironment.

[0083] To assess the model's potential response to immune checkpoint blockade therapy at the molecular level, the mechanism validation module performed target gene expression level analysis (e.g., Figure 17 H, Figure 17 I, Figure 17Total RNA was extracted from brain tumor tissue samples (as shown in SEQ ID NO: 8 and 9). The total RNA included messenger RNA of the Pdcd1 gene (encoding the PD-1 protein). Pre-defined Pdcd1 gene-specific amplification primers were obtained. Using a quantitative reverse transcription PCR detection component and the Pdcd1 gene amplification primers, the expression level of the Pdcd1 gene messenger RNA in the total RNA of the tumor tissue was accurately measured. The mechanism verification module received the measurement results and determined that: compared with the second control mouse group, the expression level of the Pdcd1 gene in the second experimental mouse group was significantly upregulated. This result indicates that MG... OE-Csmd3,BV2 It significantly altered the immunological characteristics of the tumor, providing a clear molecular mechanism basis for the subsequent direct introduction of PD-1 blockers for combination therapy.

[0084] Based on the above morphological and molecular mechanism evidence, in G422 TN -The GBM model formally conducts in vivo combined drug immunotherapy response assessment (e.g.) Figure 17 K, Figure 17 L, Figure 17 (as shown in M), re-acquired with mouse G422 TN - Primarily wild-type Kunming mice with perfectly matched GBM cell immune backgrounds were randomly divided into a second basic treatment group and a second combined treatment group. Using a stereotactic injection device, a system containing MG cells was constructed. OE-Csmd3,BV2 With mouse G422 TN A GBM cell-mediated tumor model was established using in situ co-implantation. Monoclonal antibodies targeting programmed death receptor 1 (PD-1) and corresponding isotype control antibodies were obtained. Following a pre-defined treatment protocol, the PD-1 immune checkpoint blocking antibody was intraperitoneally injected into mice in the second combination therapy group, while the isotype control antibody was intraperitoneally injected into mice in the second basic therapy group. The mechanism validation module continuously observed and received survival time data from both groups after antibody injection, obtaining the survival results of the combination therapy. The mechanism validation module used the survival results of the combination therapy to determine that the survival rate and median survival time of the second combination therapy group were significantly higher than those of the second basic therapy group. Based on the above multi-dimensional mechanism validation results, it was confirmed that MG... OE-Csmd3,BV2 It can transform immune "cold tumors" into immune "hot tumors" rich in tertiary lymphoid structures, thus possessing the core function of significantly improving the sensitivity of gliomas to PD-1 immune checkpoint blocking antibody therapy; it should be noted that the stage survival assessment is only for the anti-tumor response of the first dose, and the entire process does not involve a second challenge to the tumor.

[0085] To confirm the long-lasting and specific nature of the antitumor immune response, refer to the appendix. Figure 18This invention provides an independent method for verifying the persistence of anti-GBM immune memory. The method aims to verify whether early intervention has successfully constructed a long-term immune barrier against tumor recurrence in the host. The specific implementation steps are detailed below: First, rigorous screening and control grouping of experimental animals were conducted: This was done to obtain MG samples from the initial stage. OE-Csmd3,BV2 With G422 TN In the GBM co-implantation assessment, experimental mice that survived the initial tumor risk period and reached the pre-defined observation endpoint were formally defined as the second long-term survival mouse group. Simultaneously, a group of age- and sex-matched normal mice that had not undergone any prior tumorigenesis treatment were obtained in parallel and used as the re-challenge control group. Cell suspensions for secondary tumor re-challenge were then prepared: G422 mice stably expressing luciferase were obtained as pre-defined cell suspensions. TN GBM cells were individually placed in a pre-designed cell resuspension solution for thorough resuspension and mixing to construct a third implanted single-cell suspension specifically for in vivo tracing.

[0086] Then, the tumor rechallenge procedure was performed: the stereotactic injection device was configured and used to precisely inject the third implanted single-cell suspension into the contralateral brain tissue region of each mouse in the second long-term survival mouse group and the rechallenge control mouse group. After the tumor rechallenge treatment was completed, no anti-tumor drugs or interventions were given to mice in any group.

[0087] Then, in vivo tracking and final confirmation of immune cure are performed: using in vivo imaging detection components, in vivo images of brain tissue regions in mice after tumor rechallenge are periodically acquired to obtain secondary tumor growth data (e.g., Figure 18 As shown in B), the survival rate data of these two groups of mice were recorded simultaneously (e.g., Figure 18 As shown in C), the mechanism verification module receives and comprehensively analyzes the above data. The results clearly show that all mice in the re-challenge control group died due to rapid tumor growth in vivo. In contrast, no imaging signs of tumor growth were observed in the brains of these second batch of long-term surviving mice after tumor re-challenge, and the survival rate after re-challenge remained at 100%. Based on the above rigorous in vivo and in vitro and multi-dimensional verification results, the MG mechanism in this invention is comprehensively confirmed. OE-Csmd3,BV2 It not only inhibits the primary tumor, but also possesses the core persistent function of inducing, establishing and maintaining long-lasting and specific anti-GBM immune memory in the host.

[0088] The above embodiments are for Csmd3 + Systematic confirmation of the prognostic value and immune mechanism of microglia in the glioma microenvironment, and the Csmd3 provided in this invention. +The microglia detection method, the target subpopulation characteristic gene set, and the detection reagents targeting CSMD3 and IBA1 proteins have clinical translational value and can be substantially applied to the preparation of medical prognostic assessment products for glioma patients. Specifically, the prognostic assessment products can take the form of clinical prognostic test kits, gene-targeted sequencing panels, customized microarray chips, or supporting medical data analysis software systems. In a specific implementation and translation scheme, specific amplification primers or fluorescent probes targeting each gene sequence within the target subpopulation characteristic gene set, or CSMD3 and IBA1 antibodies as core active ingredients, can be physically encapsulated to prepare standardized clinical prognostic test kits. These kits can be directly deployed in clinical laboratories to perform rapid and standardized targeted feature detection and subpopulation abundance quantification on biopsy tissue samples or transcriptome sequencing data from glioma patients.

[0089] Furthermore, in conjunction with the "Csmd3" already clearly verified in the in-body embodiments of this invention... + The detection methods and reagents described above, which demonstrate significant sensitization to PD-1 immune checkpoint blockade antibodies and a 100% immunocure rate, can also be used to develop a clinical auxiliary assessment tool for "precisely screening patients with predisposition to immunotherapy in gliomas." In practical clinical applications, the tool operates as follows: by measuring and calculating the subgroup enrichment score and the total microglia enrichment score within the patient sample, the ratio of these two scores is calculated. This parameter is then compared to a predefined abundance threshold. If the patient is determined to be in a high-abundance sample group, clinical indications are given. Patients exhibit extremely high treatment sensitivity to immune checkpoint inhibitors such as PD-1, and after combination therapy, they are highly likely to establish long-lasting anti-GBM immune memory in vivo, making them a population with a significant advantage. Conversely, if a patient is determined to be in the low abundance sample group, it indicates that the effect of using immune checkpoint inhibitors alone may be limited, requiring timely adjustment or the adoption of alternative personalized treatment strategies. This approach not only provides precise tumor microenvironment cytological markers for predicting the long-term survival of glioma patients, but also provides solid and reliable data decision support for clinicians to formulate and optimize individualized combination immunotherapy regimens.

[0090] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A Csmd3 for assessing prognosis of a glioma patient + A microglia cell detection method characterized by, include: Acquire clinical transcriptomic data and glioma tissue samples, wherein the clinical transcriptomic data includes transcriptional sequencing data from multiple glioma patient samples; Obtain the characteristic gene set of the target subpopulation and the characteristic gene set of total microglia. The characteristic gene set of the target subpopulation is constructed based on the characteristic genes of the target cell subpopulation, which is Csmd3. + Microglia, wherein the total microglia characteristic gene set is constructed based on the characteristic genes of total microglia; Configure a single-sample gene set enrichment analysis algorithm, and use the single-sample gene set enrichment analysis algorithm to calculate the subgroup enrichment score of each glioma patient sample in the multiple glioma patient samples in the clinical transcriptome data relative to the target subgroup characteristic gene set, and the total microglia enrichment score relative to the total microglia characteristic gene set. The relative proportion value is calculated based on the target subpopulation enrichment score and the total microglia enrichment score, and the relative proportion value is determined as the clinical abundance parameter of the target cell subpopulation in each glioma patient sample. Configure a survival analysis algorithm to calculate the median value of the clinical abundance parameter, and use this as a dividing threshold to divide the multiple glioma patient samples corresponding to the clinical transcriptome data into a high abundance sample group and a low abundance sample group. Acquire clinical survival data, input the high-abundance sample group, the low-abundance sample group, and the clinical survival data into the survival analysis algorithm for calculation, and obtain survival prognosis assessment data; Using the survival prognostic assessment data, the overall survival of the high-abundance sample group and the low-abundance sample group is compared. It is determined that the overall survival of the high-abundance sample group is longer than that of the low-abundance sample group. The prognostic correlation verification results are output, showing that the clinical abundance parameters of the target cell subpopulation are positively correlated with the survival prognosis of the glioma patients corresponding to the glioma patient samples.

2. The Csmd3 for use according to claim 1 in the evaluation of the prognosis of a glioma patient. + The microglia cell detection method is characterized by comprising: The acquisition of the target subpopulation characteristic gene set and the total microglia characteristic gene set includes: Immunocured mice and control mice were constructed. Single-cell RNA was extracted from brain tumor tissue samples of the immune-cured mice and brain tumor tissue samples of the control group mice, respectively. Single-cell transcriptome sequencing was performed on the single-cell RNA to obtain single-cell transcriptome sequencing data; The Csmd3 gene was obtained from the total microglia population as a target identification gene, and the expression level of the target identification gene within the microglia population was evaluated. Based on the expression level, specific cell clusters expressing the targeted identification gene are isolated from the microglia population, and the specific cell clusters are identified as the target cell subpopulation. The characteristic genes of the target cell subpopulation are extracted, and the characteristic gene set of the target subpopulation is constructed.

3. The Csmd3 for use according to claim 2, for assessing the prognosis of a glioma patient. + The microglia cell detection method is characterized by comprising the steps of: The extraction of microglia from the single-cell transcriptome sequencing data includes: Perform quality control on the single-cell transcriptome sequencing data, remove cells whose number of expressed genes is outside the preset range and whose mitochondrial gene ratio is greater than the preset threshold, and output the quality-controlled single-cell sequencing dataset. Configure a dual-cell prediction algorithm, and use the dual-cell prediction algorithm to perform dual-cell identification processing on the quality control single-cell sequencing dataset, remove cell data with dual-cell scores greater than the threshold, and generate a standard single-cell feature matrix. Gene variability analysis is performed on the standard single-cell feature matrix to extract highly variable feature genes, and principal component analysis is performed based on the highly variable feature genes to obtain the principal component analysis results. Configure a batch effect correction algorithm and a dimensionality reduction algorithm. Use the batch effect correction algorithm to perform inter-sample batch effect correction processing on the principal component analysis results, and use the dimensionality reduction algorithm to perform data dimensionality reduction, outputting a dimensionality-reduced integrated dataset. Cluster analysis was performed on the dimensionality-reduced and integrated dataset to obtain multiple cell clusters. Differentially expressed characteristic genes of these cell clusters were calculated. A chromosome copy number variation inference algorithm was configured, and the copy number variation event data generated by the algorithm was used to annotate and divide the multiple cell clusters into G422. TN -GBM tumor cell population and non-tumor cell population, from which the total microglia population is determined, and the characteristic genes of the total microglia population are extracted to construct the total microglia characteristic gene set.

4. The Csmd3 method for assessing the prognosis of glioma patients according to claim 1. + The method for detecting microglia is characterized by... It also includes clinical histological verification steps: The glioma tissue sample was subjected to paraffin sectioning to obtain tissue section samples; Obtain the preset CSMD3 antibody and the preset IBA1 antibody; The tissue section samples were subjected to multiple immunofluorescence staining using the CSMD3 antibody and the IBA1 antibody to target and label the CSMD3 protein and IBA1 protein in the tissue section samples, thereby obtaining multiple immunofluorescence staining images. Configure an image analysis algorithm and use the image analysis algorithm to perform image feature analysis on the multiple immunofluorescence staining image to identify specific cell groups in the tissue section sample that simultaneously express CSMD3 protein and IBA1 protein, and define the specific cell groups as clinical homologous target cell groups; The spatial distribution characteristic parameters of the clinical homologous target cell population in the multiplex immunofluorescence staining image are calculated using the image analysis algorithm, and the infiltration density of the clinical homologous target cell population in the glioma tissue sample is calculated using the spatial distribution characteristic parameters.

5. The Csmd3 for use according to claim 1 in the evaluation of the prognosis of a glioma patient. + A microglia cell detection method characterized by, It also includes clinical single-cell level validation steps: Obtain the Gene Expression Omnibus (GEO) database and extract human glioma single-cell transcriptome data from the GEO database; The human glioma single-cell transcriptome data were subjected to quality control and dimensionality reduction clustering to construct a clinical single-cell sequencing dataset; Obtain preset microglia markers, and extract clinical microglia populations from the clinical single-cell sequencing dataset based on the expression characteristics of the microglia markers; The expression intensity of the clinical microglia population was evaluated using the target identification gene of the target cell subpopulation. Single cell clusters expressing the target identification gene were isolated from the clinical microglia population and defined as clinical homologous single cell subpopulations. The proportion of the clinical homologous single-cell subpopulation in the clinical microglia group is calculated. Based on the cell proportion, the presence status of the target cell subpopulation in the human clinical glioma microenvironment and its proportion in different glioma types are confirmed.

6. The Csmd3 for use according to claim 1 in the evaluation of the prognosis of a glioma patient. + A microglia cell detection method characterized by, It also includes a step of verifying the in vitro functional intervention of the target cell subpopulation: A pre-defined mouse BV2 microglia cell line was obtained as the target cell for in vitro intervention. The promoter sequence of the mouse Csmd3 gene was obtained, and a single guide RNA sequence was designed and synthesized based on the promoter sequence of the mouse Csmd3 gene. A preset CRISPR activation system is obtained, which includes a first lentiviral particle expressing an inactivated Cas9 protein, a second lentiviral particle expressing a transcription activation domain, and a single-guide RNA lentiviral backbone plasmid. The single-guide RNA sequence is cloned into the single-guide RNA lentiviral backbone plasmid to construct a single-guide RNA expression vector. A pre-defined viral packaging helper plasmid was obtained, and a third lentiviral particle carrying the single guide RNA sequence was prepared using the single guide RNA expression vector and the viral packaging helper plasmid. As a control, a pre-defined empty lentiviral particle containing a non-targeting control sequence was used instead of the third lentiviral particle. The mouse BV2 microglia were infected with the first lentiviral particles and the first resistant drug was used for initial screening to obtain a first resistant cell population expressing inactivated Cas9 protein. Then, the first resistant cell population was infected with the second lentiviral particles, and a second resistant drug was used for secondary screening to obtain a second resistant cell population that simultaneously expresses inactivated Cas9 protein and transcription activation domain. Then, the second resistant cell population is infected with the third lentiviral particles, and a final screening is performed using the third resistant drug to obtain a stable infected cell population. The first resistance drug is blast fungicide, the second resistance drug is hygromycin B, and the third resistance drug is puromycin; Transcriptional expression data and Western blot quantitative data were obtained for the stable infected cell population. Based on the comprehensive verification results of the transcriptional expression data and the Western blot quantitative data, the stable infected overexpressing cell population was identified as MG. OE-Csmd3, BV2 The stable infection control cell population was identified as MG. NC, BV2 .

7. The Csmd3 for use according to claim 6, for assessing the prognosis of a glioma patient. + The microglia cell detection method is characterized by comprising the steps of: In determining the MG OE-Csmd3,BV2 Afterwards, there is also included an in vitro tumor growth inhibition evaluation step: Obtain mouse GL261 glioblastoma (GL261-GBM) cells labeled with luciferase; Constructing an in vitro direct cell co-culture system, said MG OE-Csmd3,BV2 mixed directly with said luciferase-labeled mouse GL261-GBM cells and placed in said direct cell co-culture system for incubation in the first experimental group; Obtain a preset fluorescein substrate reagent, add the fluorescein substrate reagent to the direct cell co-culture system after incubation treatment in the first experimental group, configure a bioluminescence imaging component, use the bioluminescence imaging component to acquire bioluminescence images of the direct cell co-culture system, and extract the bioluminescence radiance value of the region of interest. An in vitro indirect cell co-culture system was constructed, and the MG cells were cultured in a pre-designed co-culture separation chamber. OE-Csmd3,BV2 The cells were physically isolated from the mouse GL261-GBM cells labeled with luciferase and seeded into the upper and lower chambers of the indirect cell co-culture system for incubation in the second experimental group. A cell viability detection component was configured, and mouse GL261-GBM cells were extracted from the lower chamber after the incubation treatment of the second experimental group. The relative cell viability data of the mouse GL261-GBM cells after indirect co-culture were measured using the cell viability detection component. Based on the bioluminescence radiance value and the relative cell viability data, the MG is determined comprehensively. OE-Csmd3,BV2 It has the ability to inhibit the proliferation of the mouse GL261-GBM cells.

8. The Csmd3 for use according to claim 6, for assessing the prognosis of a glioma patient. + The microglia cell detection method is characterized by comprising the steps of: In determining the MG OE-Csmd3,BV2 Afterwards, there is also an in vivo tumor growth inhibition evaluation step: Obtaining first GBM cells, which are mouse GL261-GBM cells with a luciferase label, and second GBM cells, which are mouse G422 TN - GBM cells; Obtain a preset cell resuspension, and then separately add the MG... OE-Csmd3,BV2 The first co-implanted cell suspension and the second co-implanted cell suspension were mixed with the first GBM cells and the second GBM cells in the cell resuspension to construct the first co-implanted cell suspension and the second co-implanted cell suspension, respectively. Experimental mice that are immune to the first GBM cells and the second GBM cells were obtained and divided into the first tumor model experimental mouse group, the first tumor model control mouse group, the second tumor model experimental mouse group, and the second tumor model control mouse group, respectively. A stereotactic injection device was configured to inject the first co-implanted cell suspension into the brain tissue region (right striatum) of the first tumor model mouse group to construct a GL261-GBM co-implanted tumor model. The second co-implanted cell suspension was then injected into the brain tissue region (right striatum) of the second tumor model mouse group to construct a G422 co-implanted tumor model. TN -GBM co-implanted tumor model; Configure in vivo imaging detection components and tissue section staining components; For the GL261-GBM co-implanted tumor model, the in vivo imaging detection component is used to acquire in vivo image data of the GL261-GBM co-implanted tumor model to obtain tumor in vivo growth data, and a portion of the brain tissue of the GL261-GBM co-implanted tumor model is extracted for hematoxylin-eosin staining. The area of ​​the largest cross section of the tumor is measured by the tissue section staining component to obtain the first tumor burden data. For the G422 TN - GBM co-implanted tumor model, directly extract its brain tissue for hematoxylin-eosin staining, measure the area of the largest cross-section of the tumor by the tissue slice staining assembly to obtain the second tumor burden data; Integrating survival time to obtain mouse survival data for each co-implanted tumor model, using the tumor in vivo growth data, the first tumor burden data, the second tumor burden data, and the mouse survival data, collectively determining the MG OE-Csmd3,BV2 Inhibiting tumor growth in an in vivo environment.

9. A Csmd3 for assessing prognosis of a glioma patient according to claim 8. + The microglia cell detection method is characterized by comprising: It also includes steps for tumor microenvironment remodeling and tertiary lymphoid structure induction analysis: Brain tumor tissue samples were extracted from the experimental mouse group and processed into paraffin sections to obtain tumor tissue sections. Obtain preset CD3 antibody and preset CD20 antibody, and use the CD3 antibody and CD20 antibody to perform immunohistochemical staining on the tumor tissue section to target and label the CD3 protein and CD20 protein in the tumor tissue section, and obtain the tumor section staining image. A microscopic image analysis algorithm is configured, and the image feature analysis of the stained tumor slice image is performed using the microscopic image analysis algorithm to identify T cells expressing CD3 protein and B cells expressing CD20 protein respectively. The cell infiltration density of the T cells and the B cells in the non-tertiary lymphoid structure region within the tumor is calculated to generate immune cell infiltration abundance data. The spatial structure of T cells expressing CD3 protein and B cells expressing CD20 protein aggregates is identified using the microscopic image analysis algorithm, and the spatial structure containing T cell and B cell aggregates is defined as a tertiary lymphoid structure. The number and area of ​​the tertiary lymphoid structures in the tumor tissue section are calculated using the aforementioned microscopic image analysis algorithm, generating structural distribution density data; It also includes an immune cell memory phenotype assessment step: Single-cell transcriptome sequencing data were obtained from immune-cured mice and control mice, and the target cell subpopulations, T cell populations and B cell populations were extracted from them. Expression analysis of key regulatory factors of innate immune memory was performed to determine that key transcription factors of innate immune memory were upregulated in the target cell subpopulation within the immune-cured mice. Based on the enrichment analysis results of the differential gene set, the significantly upregulated feature score data, and the key transcription factors of the innate immune memory, it was determined that the cellular state attribute of the target cell subpopulation in the immune-cured mice was an immune-activated state and that it had an innate immune memory phenotype. Obtain multiple memory taxonomic marker genes in T cells from a predefined set, including Cd69, Itgae, Cd44, Sell, Ccr7, Il7r, and Il15ra; The expression levels of specific T cell memory group marker genes within the T cell population were calculated. By comparing the expression levels of the specific T cell memory group marker genes in the immunized cured mice and the control mice, it was determined that the T cell population in the immunized cured mice possessed tissue residency memory phenotype characteristics, and was characterized as a tissue residency memory T cell population. Obtain preset memory B cell markers, including Cxcr5 and Cr2 genes, and analyze the expression of the memory B cell markers within the B cell population; By comparing the expression of Cxcr5 and Cr2 genes in the immune-cured mice and the control mice, it was determined that the B cell population in the immune-cured mice possesses immune memory phenotype characteristics, exhibiting memory B cell population characteristics. It also includes steps for assessing tumor-killing activation status and analyzing cell communication networks in the anti-GBM immune memory microenvironment: The single-cell transcription feature matrix is ​​extracted and input into a preset cell communication analysis algorithm for calculation and processing to obtain receptor-ligand interaction data; Based on the receptor-ligand interaction data, by comparing the immune-cured mice with the control group mice, it was determined that the overall number of intercellular interactions was increased and the tumor cell self-interactions were weakened in the immune-cured mice. The enhanced communication strength between the tissue-resident memory T cells and the tumor cells was determined, and the signaling pathways and specific receptor-ligand pairs exhibiting enhanced characteristics were extracted. The signaling pathways included PARs, TRAIL and LCK signaling pathways, and the specific receptor-ligand pairs included Cd6-Alcam, Cd226-Pvr and Tnfsf10-Tnfrsf10b. subpopulations of T cells, the plurality of T cell subpopulations comprising CD4 + tissue-resident memory T cell subpopulation and CD8 + tissue-resident memory T cell subpopulation; CD8 was identified through cell communication network analysis. + Organ-resident memory T cell subsets act as primary signal transducers and receivers, as well as CD4+. + Organ-resident memory T cell subsets act as regulators of CD8 + The main signaling agents for the activity of tissue-resident memory T cell subsets; Based on the combined analysis of the receptor-ligand interaction data, the signaling pathways, the specific receptor-ligand pairs, and the cell communication network of the T cell subpopulations, it was determined that the anti-GBM immune memory microenvironment in the immunized mice was in a state of tumor killing activation. It also includes steps for triggering an assessment of the immunotherapy response: Another batch of mice G422 TN - GBM cell immune background matched experimental mice, divided into combined treatment mouse group and basic treatment mouse group; with the MG OE-Csmd3,BV2 with the mouse G422 TN - co-implanted cell suspension of GBM cells were injected into the brain tissue region (right striatum) of the combination therapy mice group and the basic therapy mice group, respectively; Obtain a preset PD-1 immune checkpoint blocking antibody and an isotype control antibody. Inject the PD-1 immune checkpoint blocking antibody into each mouse in the combined treatment mouse group via intraperitoneal injection. Inject the isotype control antibody into each mouse in the basic treatment mouse group via intraperitoneal injection. Survival time was continuously recorded to obtain combined treatment survival data. The survival time of the combined treatment mouse group and the basic treatment mouse group was compared to determine the MG. OE-Csmd3,BV2 It has the function of inducing immunotherapy response (triggering PD-1 immune checkpoint blocking antibody response) and prolonging the survival of tumor-bearing mice; It also includes an anti-GBM immune memory assessment step: Obtain and accept the MG OE-Csmd3,BV2 With the mouse G422 TN - A group of mice that survived for a long time after co-implantation of GBM cells and reached the first preset time threshold, and age- and sex-matched normal mice were obtained as a re-challenge control group. The mouse G422 TN - the GBM cells are resuspended alone in the cell resuspension solution, building a secondary implant cell suspension; Using the stereotactic injection device, the secondary implanted cell suspension was injected into the brain tissue region (left striatum) of each mouse in the long-term surviving mouse group and the re-challenge control mouse group to perform tumor secondary implantation treatment without any additional anti-tumor treatment. Using the aforementioned live imaging detection component, live imaging of the brain tissue region (left striatum) after the secondary tumor implantation treatment is performed to obtain secondary tumor growth data; The survival time after secondary implantation was continuously recorded, and the long-term surviving mouse group whose survival time reached the second preset time threshold was defined as the immune-cured mouse group. Based on the secondary tumor growth data and the rate of long-term survival of tumor-bearing mice in the immunologically cured mouse group, the MG OE-Csmd3,BV2 with the function of establishing long-term anti-GBM immune memory.

10. The Csmd3 according to any one of claims 1 to 9 + The microglia detection method, or the target subpopulation characteristic gene set as described in claim 1, or the detection reagent targeting CSMD3 protein and IBA1 protein as described in claim 4, are used in the preparation of prognostic assessment products for glioma patients. The application also includes: Based on the prognostic correlation validation results, this study examines its application in developing an auxiliary assessment tool for screening glioma patients sensitive to immunotherapy.

Images